Seven Novel Variants of the Staphylococcal Chromosomal Cassette mec in Methicillin-Resistant Staphylococcus aureus Isolates from Ireland

Transcription

1 ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 2005, p Vol. 49, No /05/$ doi: /aac Copyright 2005, American Society for Microbiology. All Rights Reserved. Seven Novel Variants of the Staphylococcal Chromosomal Cassette mec in Methicillin-Resistant Staphylococcus aureus Isolates from Ireland Anna Shore, 1 Angela S. Rossney, 2 Conor T. Keane, 2 Mark C. Enright, 3 and David C. Coleman 1 * Microbiology Research Unit, Department of Oral Surgery, Oral Medicine and Pathology, School of Dental Science, Trinity College, University of Dublin, Ireland 1 ; National MRSA Reference Laboratory, St. James s Hospital, Dublin, Ireland 2 ; and Department of Biology and Biochemistry, University of Bath, Bath, United Kingdom 3 Received 20 September 2004/Returned for modification 16 December 2004/Accepted 28 December 2004 Methicillin-resistant Staphylococcus aureus (MRSA) isolates recovered in Irish hospitals between 1971 and 2002 were characterized using multilocus sequence typing (MLST) (n 130) and SCCmec typing (n 172). Where atypical SCCmec typing results were obtained, PCR amplification of entire SCCmec elements, analysis of amplimer mobility, and nucleotide sequencing were undertaken. MLST revealed that 129/130 isolates had the same genotypes as internationally spread MRSA clones, including ST239, ST247, ST250, ST5, ST22, ST36, and ST8. A novel genotype, ST496, was identified in one isolate. Half of the isolates (86/172) had SCCmec type I, IA, II, III, or IV. The remaining 86 isolates harbored novel SCCmec variants in three distinct genetic backgrounds: (i) 74/86 had genotype ST8 and either one of five novel SCCmec II (IIA, IIB, IIC, IID, and IIE) or one of two novel SCCmec IV (IVE and IVF) variants; (ii) 3/86 had genotype ST239 and a novel SCCmec III variant; (iii) 9/86 had a novel SCCmec I variant associated with ST250. SCCmec IVE and IVF were similar to SCCmec IVc and IVb, respectively, but differed in the region downstream of meca. The five SCCmec II variants were similar to SCCmec IVb in the region upstream of the ccr complex but otherwise were similar to SCCmec II, except for the following regions: SCCmec IIA and IID had a novel mec complex, A.4 ( meci-is1182- mecimecr1-meca-is431mec); SCCmec IIC and IIE had a novel mec complex, A.3 (IS1182- meci-mecr1-meca- IS431mec); SCCmec IID and IIE lacked pub110; SCCmec IIC and IIE lacked a region of DNA between Tn554 and the mec complex; and SCCmec IIB lacked Tn554. This study has demonstrated a hitherto-undescribed degree of diversity within SCCmec. Methicillin-resistant Staphylococcus aureus (MRSA) was first reported in Irish hospitals in years after the initial report of MRSA in England (14, 20). In line with trends worldwide, the prevalence of MRSA increased in Ireland during the 1990s, with one Irish hospital reporting a fourfold increase in the number of patients with MRSA between 1989 and 1998 (42). The European Antimicrobial Resistance Surveillance System (EARSS) reported that in Ireland in 2002, 42% of S. aureus isolates from blood cultures were methicillin resistant, one of the highest prevalence rates in Europe (8). Methicillin-susceptible S. aureus (MSSA) isolates have the potential to become methicillin resistant due to acquisition of a mobile staphylococcal chromosomal cassette (SCC) carrying the meca gene, termed SCCmec (23). The meca gene encodes an extra penicillin-binding protein (PBP) 2a or PBP 2 that has decreased affinity for -lactam antibiotics, allowing cell wall synthesis to continue despite inactivation of native PBPs (12, 35). To date, five types of SCCmec element (SCCmec I, II, III, IV, and V) and a small number of variants have been characterized (16 19, 25, 31, 32). Each SCCmec element integrates at the same site (attb scc ) at the 3 end of an open reading frame (ORF) of unknown function, designated orfx (17). * Corresponding author. Mailing address: Microbiology Research Unit, Department of Oral Surgery, Oral Medicine and Pathology, School of Dental Science and Dublin Dental Hospital, Trinity College, University of Dublin, Dublin 2, Republic of Ireland. Phone: Fax: dcoleman@dental.tcd.ie. SCCmec consists of the mec gene and cassette chromosome recombinase (ccr) gene complexes. Five classes of mec gene complex (A to E), which vary in their genetic structure, have been described (22, 24). Each mec complex consists of an intact copy of meca, a copy of IS431mec and, when present, complete or truncated mec regulatory genes meci and mecr1 (24). The ccr complex consists of the ccr genes ccra and ccrb in combination (ccrab) or ccrc alone, as well as adjacent ORFs (18, 23). The ccrab and ccrc genes encode recombinases necessary for site- and orientation-specific integration and accurate excision of the SCCmec element. Five allotypes of the ccr gene complex have been identified (16, 18, 32). The rest of the SCCmec element outside the ccr and mec complex is known as the junkyard (J) region, because it contains genes that are nonessential components of SCCmec (19). The five SCCmec types described to date are defined on the basis of the class of mec gene complex and the type of ccr complex they possess, and variants of each type are defined by the J regions. Variants of SCCmec types described include SCCmec IA, which differs from SCCmec I by the presence of an integrated plasmid, pub110, downstream of the mec complex; SCCmec IIIA, which differs from SCCmec III by the absence of pt181 and its flanking IS431 elements; and SCCmec IIIB, which lacks integrated copies of Tn554, pt181, and the mer operon with its associated insertion sequences (32). Variants of SCCmec IV include SCCmec IVa and SCCmec IVb as described by Ma and colleagues, which differ in their DNA sequences from the left 2070

2 VOL. 49, 2005 NOVEL SCCmec VARIANTS AND mec CLASSES IN IRISH MRSA 2071 extremity to the ccr complex (L-C region) but which both carry the downstream constant region (dcs) (25). Both SCCmec IVc and IVd also differ in their L-C regions (SCCmec IVc carries an integrated copy of Tn4001 flanked on either side by IS256) (19). SCCmec IV has a type 2 ccr complex and class B mec, but Oliveira and colleagues described an SCCmec IV element with type 4 ccr (25, 32). A type IVA SCCmec element has also been described that harbors the integrated plasmid pub110 (31). Multilocus sequence typing (MLST), SCCmec typing, and other molecular techniques have shown that the majority of nosocomial MRSA infections are caused by relatively few pandemic clones that have evolved by the introduction of SCCmec elements into five distinct epidemic MSSA lineages (33, 36). Using MLST and the program eburst, the five pandemic MRSA lineages can be visualized as clonal complexes (CCs), which are groups of genotypes and MLST sequence types (STs) that share a recent common ancestor (11). CCs are named after the ST of the ancestral genotype. The CCs of the major nosocomial MRSA lineages are CC5, CC8, CC22, CC30, and CC45 (10). In all of these lineages, except possibly CC22, SCCmec has been acquired on multiple occasions. All major international nosocomial MRSA clones belong to one of these five CCs. The names originally assigned to the MRSA clones represent either a unique epidemiological characteristic or signify the geographic area from which they were first isolated. A more rational and unambiguous scheme has been proposed based on ST and SCCmec type (10, 37). Using this nomenclature, some of the more common MRSA clones are named as follows: CC5, ST5-MRSA-II (New York/Japan) and ST5- MRSA-IV (Pediatric); CC8, ST239-MRSA-IIIA (Brazilian), ST239-MRSA-III (Hungarian), ST247-MRSA-IA (Iberian), ST250-MRSA-I (Archaic), ST8-MRSA-II (Irish-1), and ST8- MRSA-IV (EMRSA-2, EMRSA-6); CC22, ST22-MRSA-IV (EMRSA-15); CC30, ST36-MRSA-II (EMRSA-16); CC45, ST45-MRSA-IV (Berlin) (10, 32, 33). It has been suggested that genetic relatedness of particular MRSA isolates should be investigated by determining both the genotype of the MSSA isolate into which the SCCmec element was introduced and the type of SCCmec element it harbors (10). MLST has been shown to be the most powerful molecular technique for genotyping S. aureus isolates in long-term and global epidemiological studies (10, 33). The SCCmec type can be characterized using various PCR-based techniques that identify both the ccr and mec gene type or other sequences in the J regions specific to each SCCmec element (29, 31). The objective of the present study was the molecular characterization of representative clinical MRSA isolates recovered in Irish hospitals between 1971 and 2002, examining their genetic relatedness to each other and to internationally recognized MRSA clones using MLST and SCCmec element analysis. The results of this investigation identified seven novel SCCmec variants and two new mec classes, which were further characterized. MATERIALS AND METHODS Bacterial isolates. A total of 172 MRSA isolates representative of the most prevalent phenotypes recovered in Irish hospitals during eight study periods between 1971 and 2002 were investigated (Table 1). Isolates were chosen to include at least two isolates of each type or subtype recognized during each study period. Prior to 1988 (study periods A, B, and C), isolates were classified into four phenotypes (termed early MRSA and phenotypes I, II, and III) based on a combination of antimicrobial resistance patterns, bacteriophage typing, plasmid screening, location of resistance determinants, and Southern hybridization analysis of plasmid and chromosomal DNA (4, 5). Phenotype I and phenotype II isolates were previously identified during study period B (1976 to 1984), but only phenotype II isolates were available for the present study. Isolates recovered after 1988 (study periods D through H) were characterized by a combination of techniques that included bacteriophage typing, restriction fragment length polymorphism (RFLP) analysis of SmaI-digested total cellular DNA using pulsed field gel electrophoresis (PFGE), random amplified polymorphic DNA analysis using PCR (RAPD-PCR), plasmid screening, enterotoxin profiling, biotyping (pigment production and hydrolysis of Tween 80 and urea), and antibiogramresistogram (AR) typing (using a panel of 22 antimicrobials [the panel was extended to 23 in 1998]) (39 42, 44). In study periods D and E, isolates were assigned an AR type on the basis of the pattern of resistance to the antimicrobials in the AR typing panel in conjunction with additional data obtained from the phenotypic and genotypic investigations described above (39 41). Forty-four AR types (AR01 to AR44) and a number of subtypes have been recognized to date. Isolates recovered between 1971 and 1998 were from Dublin hospitals only. Isolates from 1999 were recovered from patients in hospitals that participated in the North/South Study of MRSA in Ireland, a 2-week prevalence study of MRSA in both the Republic of Ireland and the North of Ireland (3, 44). Isolates of each AR type and subtype from both the North and South of Ireland were included. Ninety-eight percent of Irish hospitals participated in the North/South Study. The isolates recovered in 2002 were from blood culture specimens from those Irish hospitals (n 23) that participated in EARSS (43). This level of participation in EARSS represents a population cover rate of 90%. Details of study isolates are shown in Table 1. Among isolates from 1999, a previously unfamiliar AR pattern, designated AR43, was predominant among isolates from the North of Ireland. AR43 isolates from six patients were recovered in mixed culture with MRSA isolates exhibiting other AR patterns (five AR13 and one AR14 pattern) (3, 44). Proportionally more isolates exhibiting AR patterns AR43, AR13, and AR14 were included in the present study to investigate this observation. Bacterial storage and culture conditions. MRSA isolates were stored at 80 C in Protect bacterial preserver vials (Technical Services Consultants Ltd., Heywood, United Kingdom) and were routinely cultured onto trypticase soy agar (Oxoid Limited, Basingstoke, United Kingdom) prior to incubation overnight at 37 C. DNA isolation and amplification. Chromosomal DNA was prepared from each isolate using the QIAGEN DNeasy kit system (QIAGEN, Crawley, West Sussex, United Kingdom) according to the manufacturer s instructions. DNA was amplified in a Thermo Hybaid Multiblock system thermal cycler (Thermo Hybaid, Ashford, Middlesex, United Kingdom). MLST. MLST was performed on 130 isolates (Table 1) by PCR amplification of internal fragments of seven housekeeping genes by using a previously described procedure and primers (9). Sequencing of both DNA strands was performed using the ABI Big-Dye Fluorescent Terminator system and an ABI 3700 automated sequencer (Applied Biosystems, Warrington, United Kingdom) at the Genomics Facility at the University of Bath (Bath, United Kingdom). The alleles at each of the seven housekeeping loci were identified by comparing the sequences obtained from the test isolates with sequences held in the MLST database ( This database was also used to identify the allelic profile and hence the ST of each isolate. SCCmec typing. SCCmec elements from all 172 isolates included in this study were typed using two methods. Firstly, a previously reported method (29) described in the present study as the simplex method was used to amplify the cassette chromosome recombinase (ccr) and mec gene complexes (ccr-mec genes). The primers of Ito et al. and Robinson and Enright were used for amplification of the ccr and mec genes, respectively (17, 37). Secondly, the multiplex PCR method of Oliveira and de Lencastre, which utilizes primers designed specifically to distinguish different regions of each SCCmec type, was used as described previously (31). Nucleotide sequencing of SCCmec. The entire nucleotide sequence of the SCCmec element from two isolates was determined. One isolate (AR13.1/ ) from study period G (1999) was chosen because it was representative of isolates with ccr-mec genes indicative of SCCmec II by the simplex PCR method but lacking one to three of the five amplimers typical of SCCmec II when analyzed by the multiplex PCR SCCmec typing method. The second isolate (AR43/3330.1) from study period G also was chosen because it was representative of isolates with ccr-mec genes typical of SCCmec IV by the simplex method

3 2072 SHORE ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 1. Irish MRSA isolates investigated from each study period and typed by MLST and SCCmec typing Isolate phenotype a No. of isolates investigated in study period b : No. of isolates typed by: A ( ) B ( ) C ( ) D (1989) E ( ) F (1998) G (1999) H (2002) MLST SCCmec typing Early MRSA Phenotype II Phenotype III AR AR AR AR AR07.0/ AR07.3/ AR AR AR AR AR AR AR AR AR ARNT (urea ve) c New01 d New02 d New03 d Unf01 e Total in this study Total in original study NA f NA a The most prevalent phenotypes during each study period are underlined., isolates exhibiting this phenotype were not recovered during the study period. b Isolates from groups A to F were from Dublin; isolates in group G were from the North and South of Ireland; isolates in group H were from the South of Ireland. References for the original study with each group are as follows: group A (reference 5); B (5); C (4); D (39, 42); E (39, 42); F (42); G (3, 44); H (38, 43). c No AR type could be assigned. PFGE analysis showed that most urease-positive isolates exhibiting the AR06 pattern yielded PFGE patterns unlike the patterns obtained from other AR06 isolates. Hence, assigning urease-positive AR06 isolates to AR type 06 was considered incorrect, and these isolates were assigned to no type (NT) pending PFGE, MLST, and SCCmec analysis. d These isolates were not assigned an AR type because they produced unfamiliar AR patterns; the AR type is pending on the results of plasmid screening, enterotoxin profiling, PFGE, phage typing, biotyping, MLST, and SCCmec typing. e This isolate yielded an unfamiliar AR pattern and was not assigned an AR type, pending the outcome of MLST and SCCmec typing results in this study. f NA, not applicable. but with one of the two amplimers characteristic of SCCmec IV missing when isolates were investigated by the multiplex method. Overlapping primers were designed to amplify the entire SCCmec element from both isolates using previously published SCCmec II and SCCmec IVa nucleotide sequences (17, 25) obtained from the Entrez-PubMed Nucleotide database website ( search &DB nucleotide) (accession numbers D86934 [SCCmec II] and AB [SCCmec IVa]). The primers used to amplify and sequence the SCCmec elements are listed in Table 2. DNA fragments for sequencing and cloning were obtained by PCR amplification of chromosomal DNA using either Pfu DNA polymerase (Promega) for primer pairs dcs F and orfx R (Table 2) or the Expand Long Template PCR system (Roche) for all other primer sets (Table 2) according to the manufacturers instructions. When each primer pair was used to amplify DNA from isolate AR13.1/3330.2, template DNA from at least one isolate exhibiting SCCmec II by both SCCmec typing methods was included as a control. Similarly, for AR43/ template DNA from at least one control isolate exhibiting SCCmec IV by both methods was included. Amplimers were purified using either the GenElute PCR Clean-Up kit (Sigma-Aldrich) or Wizard SV Gel and PCR Clean-Up system (Promega) prior to direct sequencing (Lark Technologies, Essex, United Kingdom) using primer walking or cloning into pbluescript II KS( ) phagemid and sequencing. Ligations of PCR products to pbluescript were facilitated by the presence of appropriate restriction endonuclease cleavage sites within the primers (Table 2). Transformation of competent Escherichia coli DH5 prepared using CaCl 2 and identification of recombinants using blue-white selection were performed according to the methods of Sambrook and Russell (45). DNA was sequenced using an Applied Biosystems 373A DNA sequencer (Foster City, Calif.) and dye-labeled terminators. Analysis of chromatograms and sequences were carried out using the 373A Data Analysis software program version (Applied Biosystems) and DNA Strider 1.3f11 software (CEA/Saclay, Gif-sur- Yvette, France), respectively. Homology searches were performed using BLAST ( The CLUSTAL W sequence alignment computer program was used for alignments of nucleotide sequences (13). Investigation of variant SCCmec elements. The SCCmec elements harbored by all isolates identified as having SCCmec types II and IV ccr-mec genes by the simplex method but with unusual SCCmec multiplex patterns were investigated by PCR amplification. The primers used for amplification and sequencing of SCCmec from isolate AR13.1/ (representative of isolates with SCCmec II ccr-mec genes by the simplex method but lacking one to three of the five SCCmec II multiplex amplimers) were used on all atypical SCCmec II isolates (Table 2). Similarly, the primers for amplification and sequencing of the SCCmec element of isolate AR43/ (representative of isolates with SCCmec IV ccr-mec genes by the simplex method but lacking one SCCmec IV multiplex amplimer) were used on all atypical SCCmec IV isolates (Table 2). PCR products were separated by electrophoresis in 1% (wt/vol) agarose gels, and the sizes of amplimers were compared to those obtained by amplification using the same primers on template DNA from AR13.1/ and AR43/ and to previously published SCCmec element sequences. Amplimers that showed variation from expected sizes were also sequenced. Nucleotide sequence accession numbers. Nucleotide sequences were submitted to the GenBank database under accession numbers AJ (SCCmec IIE from MRSA isolate AR13.1/3330.2), AJ (SCCmec IVE from MRSA isolate AR43/3330.2), AJ (partial mec complex A.4), and AJ (partial SCCmec IIB). RESULTS MLST. Of the 172 MRSA isolates investigated in this study, 130 were examined by MLST (Table 1). These 130 isolates

4 VOL. 49, 2005 NOVEL SCCmec VARIANTS AND mec CLASSES IN IRISH MRSA 2073 TABLE 2. Primers used in this study for SCCmec element amplification and sequencing Isolate Primer pair Nucleotide sequence g Restriction site Nucleotide coordinates SCCmec region amplified AR13.1/ IRLII F CCCTCGAGGGCTCTGCGTATCAGTTAATGA XhoI a L-C c ccra R ATTTGCGGCCGCGCTTCGATAGCCTGTTTCTG NotI a ccra F ACGCGTCGACCAAGTCATAGGCTATTTACG SalI a C-M d (ccra2-tn554) Tn554 R ACGCGTCGACAAGCTATCCACGTTCAATCTCAAC SalI a Tn554 F ATTTGCGGCCGCCTTTAAAGGGTTTCGGAATA NotI a C-M d (Tn554-mecR1) mecr1 R CCCTCGAGGGCAATGCCTAAACCTAATCG XhoI a mecr1 F ATTTGCGGCCGCCGATTAGGTTTAGGCATTG NotI a M-I e (mecr1-ugpq) ugpq R CCCTCGAGGGGCTTCTGCAGGATCTTGG XhoI a ugpq F CCCTCGAGGGCCAAGATCCTGCAGAAGC XhoI a I-R f (ugpq-dcs) dcs R ATTTGCGGCCGCCGGTCATGGCTATGATTTAG NotI a dcs F ATTTGCGGCCGCGTCAATGAGATCATCTACAT NotI a I-R f (dcs-right SCCmec junction) orfx R ATTTGCGGCCGCCCCAAGGGCAAAGCGAC NotI a AR43/ IRLII F CCCTCGAGGGCTCTGCGTATCAGTTAATGA XhoI a L-C c ccra R ATTTGCGGCCGCGCTTCGATAGCCTGTTTCTG NotI a ccra F ATTTGCGGCCGCCAAGTCATAGGCTATTTACG NotI b C-M d (ccra2-meca) meca R ACGCGTCGACCCTATCTCATATGCTG SalI b meca F ATTTGCGGCCGCGATTGGGATCATAGCGTCAT NotI b M-I e (meca-is431mec) ISmec R ACGCGTCGACACGGTGATCTTGCTCAATGA SalI b ISmec F CCCTCGAGGGCCTGACTGTCATTGTAC XhoI b I-R f (IS431mec-right SCCmec junction) orfx R ATTTGCGGCCGCCCCAAGGGCAAAGCGAC NotI b a Nucleotide coordinates from SCCmec type II, accession number D b Nucleotide coordinates form SCCmec type IVa, accession number AB c L-C, the region from the left chromosome-sccmec junction to the beginning of the ccr complex. d C-M, the region from the ccr complex to the mec complex. e M-I, the region from the mec complex region to IS431mec. f I-R, the region from IS431mec to the right extremity of SCCmec. g Text in bold indicates primer sequences, text in italics indicates restriction sites added to primer sequences, and plain text indicates extra bases added to aid with cloning. included at least one isolate from each of the most common phenotypes identified among MRSA recovered between 1971 and 2002 from Irish hospitals (Table 1, study periods A to H). Multiple isolates were included for the predominant phenotypes during each study period (Table 1), except for study period B (1976 to 1984), for which isolates from only one of the two predominant phenotypes were available. MLST identified 10 STs belonging to each of the five major CCs described to date and one singleton (ST12) among the MRSA isolates investigated (Table 3). Isolates recovered prior to 1985 exhibited ST250, while the STs of the two most prevalent strains in 1989 were ST250 and ST239 (Tables 1 and 3). ST8 predominated in the 1990s, but by the late 1990s the prevalence of isolates with ST36 and ST22 increased (Tables 1 and 3). By 2002, ST22 had become the major ST. All except one ST were identical to those previously described. The exception was a double locus variant of ST5 (ST496), which differed from ST5 at both the arcc and yqil alleles. ST5 has alleles 1 and 10 of arcc and yqil, respectively, whereas ST496 had a new allele, 63, of arcc and allele 28 of yqil. The arcc alleles 63 and 1 differ at a single nucleotide site, as do alleles 10 and 28 of yqil. This novel ST was obtained from one isolate from 2002 exhibiting AR type AR07.2. SCCmec typing. Using both the simplex and multiplex SCCmec typing methods 50% (86/172) of isolates carried SCCmec type I, IA, II, III, or IV (Table 3). SCCmec I was identified in 10 isolates (5.8%), SCCmec II was identified in 30 isolates (17.4%), SCCmec III was found in nine isolates (5.2%), and SCCmec IV was identified in 34 isolates (19.8%). Three isolates (1.7%) were found to harbor SCCmec I by the simplex method, but the multiplex method showed that they carried the previously described variant element SCCmec Ia (Table 3). Identification of variant SCCmec types. Eighty-six isolates (50%) harbored two apparently different SCCmec elements when tested by both SCCmec typing methods, or the SCCmec type could not be inferred because the multiplex patterns produced did not exactly match those of previously described SCCmec elements (Table 3; Fig. 1). Nine isolates were found to carry SCCmec I by the simplex method but failed to yield the 495-bp band characteristic of SCCmec I elements by the multiplex method, which results from amplification of a region downstream of the gene encoding a plasmin-sensitive surface protein (pls) found in SCCmec I (Table 3). Furthermore, the multiplex pattern obtained was indistinguishable from SCCmec IV (Fig. 1A, lanes 2 to 5). This pattern was designated SCCmec I pls. Three isolates were identified by the simplex method as carrying SCCmec III but their multiplex pattern, although similar to that of SCCmec III, was missing a band at 243 bp (Table 3; Fig. 1A, lanes 7 to 9). This band is the product of amplification of the locus between pi258 and Tn554 and is found in all SCCmec III elements characterized to date. This pattern was designated SCCmec III pi258/tn554. All 20 AR43 isolates were classified as SCCmec type IV by the simplex method but gave a multiplex PCR pattern from which the SCCmec type could not be inferred because only the multiplex meca control amplimer found in all SCCmec types was obtained (Table 3; Fig. 1B, lanes 7 to 9). This multiplex pattern could be interpreted as a variant SCCmec IV multiplex pattern missing the 342-bp amplimer, which results from the amplification of dcs found in SCCmec IV. This pattern was called SCCmec IV dcs. The remaining 54 isolates belonged to phenotypes AR05, AR13, AR14, New01, and New03 (Table 1) and were identi-

5 2074 SHORE ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 3. MLST, SCCmec type, and variant SCCmec elements of MRSA recovered in Ireland, 1971 to 2002 SCCmec type based on: CC ST AR type or phenotype a Simplex PCR Multiplex PCR b SCCmec name c Variant SCCmec elements AR01, AR15, AR23, AR44 III III 239 PhIII, AR09 III III pi258/tn Early MRSA (n 1), PhII I I (n 10) (n 4) 250 PhII (n 1), AR02 (n 3) I IVorI pls (n 9) 247 AR22, New02 I Ia 8 AR05 II II kdp IIB SCCmec IVb L-C region and lacks Tn554 f 8 AR13 (n 6), AR14 (n 3) II II kdp IIA SCCmec IVb L-C region; class A.4 mec d, f 8 AR14 (n 10) II II kdp & meci IIC SCCmec IVb L-C region; class A.3 mec d ; lacks ORFs between Tn554 and mec complex f 8 AR13 (n 19), AR14 (n 3), New01, New03 II II kdp & pub110 IID SCCmec IVb L-C region; class A.4 mec d ; lacks pub110 and second copy of IS431 f 8 AR13 (n 1), AR14 (n 6) II IV or II kdp, meci & pub110 IIE fied as SCCmec II by the simplex method but gave variant SCCmec II multiplex patterns that lacked one to three of the five amplimers that would identify them as harboring SCCmec II (Table 3). Ten isolates lacked a 284-bp band corresponding to the amplification product from a region of the kdp operon found in SCCmec II (Table 3). This pattern was designated SCCmec II kdp (Fig. 1B, lane 2). Ten isolates lacked both the 284-bp band (kdp operon) and a 209-bp amplimer (Table 3). The latter results from amplification of part of the meci gene in SCCmec II. This pattern was termed SCCmec II kdp & meci (Fig. 1B, lane 3). Twenty-seven isolates lacked the 284-bp band (kdp operon) and the 381-bp amplimer (Table 3). This 381-bp amplimer results from amplification of the left junction between IS431mec and pub110 found in SCCmec II. This pattern was called SCCmec II kdp & pub110 (Fig. 1B, lanes 4). Seven isolates produced a multiplex PCR pattern indistinguishable from the pattern of SCCmec IV (Table 3). This pattern could also be interpreted as a SCCmec II pattern missing the 284-bp (kdp operon), 209-bp (meci gene), and 381-bp (pub110) amplimers. This pattern was designated SCCmec II kdp, meci & pub110 (Fig. 1B, lane 5). Investigation of the SCCmec elements harbored by ST8 isolates with unusual SCCmec II and SCCmec IV multiplex patterns. Since all ST8 isolates (n 74) exhibited variant multiplex patterns (Table 3), we decided to further investigate these SCCmec IVb L-C region; class A.3 mec d ; lacks pub110, second copy of IS431, and ORFs between Tn554 and mec complex f 8 AR43 (n 17) IV IV dcs IVE SCCmec IVc variant, lacks dcs; has an unusual sequence in I-R region 8 AR43 (n 3) IV IV dcs IVF SCCmec IVb variant, lacks dcs; has an unusual sequence in I-R region 5 5 AR07.3/07.4, AR11 II II 5 Unf01 IV IV 496 AR07.2 (n 1) II II AR07.0/07.2 (n 6) II II 30 ARNT (n 4) IV IV AR06 IV IV ARNT (n 1) IV IV S e 12 ARNT (n 1) IV IV a Where more than one ST/SCCmec type was identified within a phenotype (e.g., AR13 or AR07.0/07.2), the numbers represented by each ST or SCCmec type are shown in parentheses after the AR type or phenotype. For isolates identified as having genotype ST250, the numbers of isolates analyzed by MLST differed from the numbers subjected to SCCmec typing (TABLE 1), and more than one SCCmec type was identified within this genotype; therefore, the numbers of isolates identified with each ST and SCCmec type are indicated in parentheses after the phenotype and SCCmec type, respectively. Otherwise, numbers of isolates are as indicated in TABLE 1. b Some patterns obtained with the multiplex PCR SCCmec typing method lacked amplimers corresponding to the amplification of the determinants indicated. c SCCmec name assigned during the present study. d These two variants of mec complex class A (designated A.3 and A.4) were recognized during the present study (Fig. 2 and 3). e Singleton, no ancestral genotype has been assigned. f SCCmec II variant. SCCmec elements. Firstly, the entire SCCmec element from two ST8 isolates, AR 13.1/ and AR43/3330.1, with variant SCCmec II and SCCmec IV multiplex patterns, respectively, were sequenced. Isolate AR13.1/ was representative of isolates with SCCmec II by the simplex method but with SCCmec multiplex patterns lacking one to three of the amplimers typical of an SCCmec II element. AR 13.1/ exhibited multiplex pattern SCCmec II kdp, meci & pub110 (i.e., it lacked all three bands, kdp, meci, and pub110). Isolate AR43/ was representative of isolates with SCCmec IV by the simplex method but with an SCCmec multiplex pattern lacking the 342-bp dcs amplimer (SCCmec IV dcs). The genomic organizations of these two variant SCCmec elements based on their nucleotide sequence are shown schematically in Fig. 2. The genomic structures of the SCCmec elements of the remaining 72 ST8 isolates were also determined. Novel variant of SCCmec II identified in AR13.1/ The SCCmec element in AR13.1/ consisted of a 27-kb sequence containing a combination of regions previously identified in SCCmec II and IV and a novel mec complex (Fig. 2). Amplification of the L-C region yielded an amplimer of 5.6 kb, which is similar in size to the expected amplimer for SCCmec IVb (Table 4). The nucleotide sequence of this 5.6-kb amplimer was found to have 99.7% similarity to the L-C region

6 VOL. 49, 2005 NOVEL SCCmec VARIANTS AND mec CLASSES IN IRISH MRSA 2075 Downloaded from FIG. 1. SCCmec multiplex patterns and variant multiplex patterns found in Irish MRSA isolates. (A) Lane 1, SCCmec type I multiplex pattern; lanes 2 to 5, SCCmec IV multiplex pattern produced by AR02 isolates and six phenotype II isolates with SCCmec type I ccr-mec genes by the simplex method; lane 6, SCCmec type III multiplex pattern; lanes 7 to 9, the variant SCCmec III multiplex pattern (III pi258/tn554) characterized by the absence of the pi258/tn bp amplimer, produced by AR09 and phenotype III isolates with SCCmec type III ccr-mec genes by the simplex method. (B) Lane 1, SCCmec type II multiplex pattern; lanes 2 to 5, the SCCmec multiplex patterns produced by isolates with phenotypes AR05, AR13, AR14, New01, and New03 with type II ccr-mec genes by the simplex method. These multiplex patterns appear to be related to the type II SCCmec multiplex pattern but lack the kdp amplimer (284 bp) (II kdp, lane 2), or lack the kdp and meci (209-bp) amplimers (II kdp & meci, lane 3), or lack the kdp and pub110 (381-bp) amplimers (II kdp & pub110, lane 4), or lack the kdp, meci, and pub110 amplimers (II kdp, meci & pub110, lane 5). This last pattern is the same as SCCmec IV. Lane 6, SCCmec type IV multiplex pattern; lanes 7 to 9, multiplex pattern produced by isolates with the AR43 phenotype (IV dcs), which consists of the meca control amplimer only. Lane L, 100-bp DNA ladders used as molecular size reference markers. on April 26, 2018 by guest of SCCmec IVb and less than 10% similarity to the L-C regions of SCCmec types I, II, III, IVa, and IVc. Analysis of the sequence of the ccr complex identified type 2 ccr genes (ccra2 and ccrb2) with closest similarity to those of SCCmec II and SCCmec IV elements (Table 5). Amplification and sequencing of the region between the ccr and mec complexes (C-M) in AR13.1/ using two primer pairs yielded a 7-kb and 6.5-kb product (Table 4) and confirmed the presence of a truncated integrated transposon, Tn554 (Table 5). Five of the six ORFs (tnpa, tnpb, tnpc, spc, and erma) previously identified within Tn554 of SCCmec II were present in the SCCmec element of AR13.1/3330.2, and each ORF showed 100% homology at the amino acid level with those of SCCmec II (Table 5). However, Tn554 of SCCmec of AR13.1/ was truncated by 1,000 bp, resulting in the absence of an ORF of unknown function in Tn554. The six ORFs spanning a 5-kb region encoding hypothetical proteins usually found between Tn554 and the mec complex in SCCmec II were also absent (Fig. 2). Instead of these ORFs, a 1,865-bp IS1182 sequence was present. A 253-bp truncated version of meci ( meci) was identified (meci gene is usually 372 bp) (Fig. 2). Sequences of mecr1 and meca from AR13.1/ showed

7 2076 SHORE ET AL. ANTIMICROB. AGENTS CHEMOTHER. FIG. 2. Schematic diagram showing the genomic organization of SCCmec II, IIA, IIB, IIC, IID, IIE, IVa, IVb, IVc, IVE, and IVF elements. SCCmec IIA, IIB, IIC, IID, IIE, IVE, and IVF are novel variants that were identified in the present study. SCCmec II, IVa, IVb, and IVc were identified previously (19, 25), and the organization was determined based on the nucleotide sequences in the GenBank database under accession numbers D86934, AB063172, AB063173, and AB096217, respectively. The structure of SCCmec IIE and IVE were determined by sequencing of the entire SCCmec element of Irish MRSA isolates AR13.1/ (IIE; accession number AJ810120) and AR43/ (IVE; accession number AJ810121). The structures of SCCmec IIA, IIB, IIC, IID, and IVF were determined using the primers shown in Table 3 by observing the mobility of amplimers in agarose gels and in some cases by sequencing.

8 VOL. 49, 2005 NOVEL SCCmec VARIANTS AND mec CLASSES IN IRISH MRSA 2077 TABLE 4. Correlation between SCCmec multiplex types and amplimer sizes obtained using overlapping primer pairs to amplify the novel SCCmec element variants and comparison to those expected from typical SCCmec II and SCCmec IV elements SCCmec multiplex type SCCmec element L-C, IRLII F-ccrA R C-M, ccra F-Tn554 R C-M, Tn554 F-mecR1 R Amplimer size a (kb) with primer pair M-I, mecr1 F-ugpQ R I-R, ugpq F-dcs R I-R, dcs F-orfX R C-M, ccra M-I, meca I-R, ISmec F-mecA R F-ISmec R F-orfX R Reference II II II kdp IIA * * This study II kdp IIB No amplimer c This study II kdp & meci IIC This study II kdp & pub110 IID This study II kdp, meci & IIE b 5.6* 7* 6.5* 4.1* 2.2* 1.7* This study pub110 IV IVa 8.9 No amplimer d No amplimer d IV IVb 5.6 No amplimer d No amplimer d IV IVc 6.1 No amplimer d No amplimer d 4.1 No amplimer e No amplimer e IV dcs IVE b 6.1* No amplimer d No amplimer d 4.1 No amplimer e No amplimer e 8.5* 4.7* 4.7* This study IV dcs IVF 5.6* No amplimer d No amplimer d 4.1 No amplimer e No amplimer e * This study a Amplimer sizes for SCCmec IIA, IIB, IIC, IID, IIE, IVE, and IVF were calculated either by sequencing (indicated by *) or based on electrophoretic mobility in agarose gels. Amplimer sizes for SCCmec II, IVa, IVb, and IVc were calculated from sequences in the GenBank database, accession numbers D86934, AB063172, AB063173, and AB096217, respectively. All amplimer sizes are approximate. L-C, C-M, M-I, and I-R indicate the region of the SCCmec element amplified. b SCCmec elements IIE and IVE were sequenced in full from isolates AR13.1/ and AR43/3330.1, respectively. c No amplimer due to no Tn554; primers ccra F-mecR1 R were used instead, resulting in an 11.5-kb amplimer, which was sequenced. d No amplimer due to no Tn554 in SCCmec IVa, IVb, IVc, IVE, and IVF. e No amplimer due to no dcs region in SCCmec IVc, IVE, and IVF. 100% homology to sequences of other SCCmec elements (Table 5). Amplification of the region across IS431mec to the dcs region downstream of the mec complex yielded a product of 2.2 kb, which is smaller than that of any previously described SCCmec element including SCCmec II and IVb (Table 4). Sequence data from this product showed that the hypervariable (HV) region (located between meca and IS431mec) of AR 13.1/ is 997 bp shorter than that of SCCmec II and SCCmec IVb. The IS431mec region was 100% homologous to that of other SCCmec elements but lacked all DNA sequences associated with the integrated plasmid pub110 or the second IS431 sequence found in SCCmec II (Fig. 2). The extreme right of the SCCmec element of AR13.1/ from IS431 to the right chromosomal-sccmec junction (I- R region) showed 100% homology with the same region of SCCmec II (Table 5). The 15-bp direct repeat sequence characteristic of SCCmec elements was identified at the right extremity of SCCmec (DR-R) and a 26-bp inverted repeat sequence (IR-R) that was identical to that of SCCmec IVb and SCCmec II was also found. All SCCmec elements identified to date have been found to be integrated at exactly the same nucleotide position of orfx. This ORF was identified outside the IR-R region, and SCCmec of AR13.1/ was integrated at the same nucleotide position in orfx as other SCCmec elements. The novel genomic structure of the mec complex of the SCCmec element of AR13.1/ (IS1182- meci-mecr1- meca-is431mec) was more similar to the class A mec complex of MRSA (meci-mecr1-meca-is431mec) found in SCCmec II and SCCmec III than to the class B mec complex (IS1272- mecr1-meca-is431mec) of SCCmec I and IV (Fig. 3). Due to its similarity to the class A mec complexes, this novel mec complex was designated class A.3 mec (Fig. 3). Novel variant of SCCmec IV identified in AR43/ Isolate AR43/ carried a 23-kb SCCmec element, which is within the size range (20 to 25 kb) of previously described SCCmec IV elements, but sequence data indicated that it is distinctly different. The genomic structure of this SCCmec element showed most similarity to SCCmec IVc (Fig. 2). Amplification of the L-C region of AR43/ yielded an amplimer of ca 6.1 kb (Table 4), and sequencing revealed that it had 100% similarity with the L-C region of SCCmec IVc (Table 6). Amplification of the region from the ccr complex to IS431mec yielded amplimers that showed almost 100% sequence identity with the same regions of SCCmec IVc (Table 6). A class B mec complex was identified (Fig. 2), and an I-R region different from any previously described SCCmec element was determined. An amplimer of 4.7 kb was obtained following amplification of this region (Table 4). Sequence analysis identified IS431mec, but a 4.4-kb region was identified between it and the end of orfx. Searches of the GenBank database showed that 1.2 kb of this sequence had 100% homology to a sequence of the right extremity of the SCCmec element of two other MRSA isolates (15). This sequence in the GenBank database consists of orfx and the extreme right region of the SCCmec element, including the direct repeat (DR-R) and inverted repeat (IR-R) regions. The rest of the sequence between IS431mec and IR-R identified in AR43/ was found to have no homology to any sequences in the GenBank database. Determination of the genomic organization of the SCCmec elements of the remaining ST8 isolates with unusual SCCmec II and SCCmec IV multiplex patterns. In total, 74 isolates had the ST8 genotype and gave variant SCCmec II and IV multiplex patterns (Table 3). In addition to determining the nucleotide structure of the SCCmec elements of two of these isolates as described above, the SCCmec elements of the remaining 72 isolates were characterized by analyzing the mobility of amplimers on agarose gels obtained with the overlapping primers used to amplify and sequence the SCCmec from isolates AR13.1/ and AR43/ In some cases, sequencing was performed on amplimers that differed in size from that expected from previously described SCCmec elements. These results are summarized in Table 3 and Table 4. This analysis identified five variant SCCmec elements, in addition to the two identified above by complete nucleotide sequencing. Of the

9 2078 SHORE ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 5. Identities between ORFs of the novel SCCmec II variant element of Irish MRSA isolate AR13.1/ and the conventional SCCmec IVb and II elements SCCmec element and ORF Location a ORF b Homology to ORFs of SCCmec IVb % Homology c Gene product ORF Homology to ORFs of SCCmec II b % Homology c Gene product IIE CM Hypothetical protein IIE CM Hypothetical protein IIE M Hypothetical protein IIE M Conserved hypothetical protein N Conserved hypothetical protein IIE05 (ccra2) M Cassette chromosome recombinase A2 N Cassette chromosome recombinase A2 IIE06 (ccrb2) M Cassette chromosome recombinase B2 N Cassette chromosome recombinase B2 IIE M Conserved hypothetical protein N Conserved hypothetical protein IIE M Conserved hypothetical protein N Conserved hypothetical protein IIE M Conserved hypothetical protein N Conserved hypothetical protein IIE d N Pseudogene IIE11 (tnpa) N Transposase A IIE12 (tnpb) N Transposase B IIE13 (tnpc) N Transposase C IIE14 (spc) N O-Nucleotidyltransferase IIE15 (erma) CN rrna adenine N-6-methyltransferase IIE16 (tnp IS1182) IIE17 ( meci) CN Methicillin resistance protein MecI IIE18 (mecr1) CM Truncated signal transducer CN Signal transducer protein MecRI protein MecR1 IIE19 (meca) M Penicillin binding protein 2a N Penicillin binding protein 2a IIE CM Conserved hypothetical protein CN Conserved hypothetical protein IIE21 (ugpq) CM Glycerophosphoryldiester phosphodiesterase N Glycerophosphoryldiester phosphodiesterase IIE22 (tnp IS431mec) M Transposase for IS431mec N Transposase for IS431mec IIE CM Conserved hypothetical protein CN Conserved hypothetical protein IIE CM Truncated conserved hypothetical protein CN Truncated conserved hypothetical protein a The nucleotide positions were determined based on the sequences for SCCmec IIE deposited in the GenBank database under accession number AJ b ORFs and gene products from SCCmec II and SCCmec IVb, based on nucleotide sequences in the GenBank database, accession number D86934 and AB063173, respectively (17, 25). c Based on amino acid identity. d, no significant homology. seven variants recognized, five SCCmec II variants (designated SCCmec IIA, IIB, IIC, IID, and IIE) were identified among the 54 isolates with variant SCCmec II multiplex patterns, and two novel SCCmec IV variants (designated SCCmec IVE and IVF) were found among the 20 isolates with variant SCCmec IV multiplex patterns (Table 3). A schematic representation of all seven SCCmec variants is shown in Fig. 2. Additional novel variants of SCCmec II. Significant size variation was observed in the mobility of amplimers obtained from variant SCCmec II elements compared to the sizes expected from isolates with typical SCCmec II when the following three regions were amplified: the L-C region, Tn554 to the mec complex (part of C-M region), and across pub110 (part of I-R region) (Table 4). The 54 SCCmec II variants yielded the 5.6-kb fragment characteristic of the L-C region of SCCmec IVb (Table 4; Fig. 2). No amplimer obtained following amplification of the region from Tn554 to mecr1 in the SCCmec II variants matched the size expected from a typical SCCmec II element (Table 4). All but one isolate with the multiplex pattern SCCmec II kdp (n 9) and all isolates with the multiplex pattern SCCmec II kdp & pub110 (n 27) produced amplimers of ca. 13 kb. This 13-kb region was sequenced from one isolate, and it was found that IS1182 was inserted within the meci gene, near the 3 end, and that meci carried a 16-bp deletion (Fig. 3). This new mec class A complex consisting of meci-is1182- meci-mecr1-meca-is431mec was designated mec class A.4 (Fig. 3). The Tn554-mecR1 region of one isolate with the multiplex pattern SCCmec II kdp could not be amplified with the same primers used on the other SCCmec II variants. Instead, a primer pair was used to amplify the DNA region from the ccr complex to mecr1, resulting in a ca kb amplimer (Table 4). Sequencing data revealed that this isolate retained only a few bases of the transposon Tn554 normally found in SCCmec II and that all Tn554 ORFs were absent. However, the six ORFs usually found between Tn554 and the mec complex of SCCmec II were all present. This isolate also harbored a class A mec complex. Amplification of the same region in isolates with the multiplex pattern SCCmec II kdp & meci (n 10) and SCCmec II kdp, meci & pub110/iv (n 7) yielded an amplimer of the same size as the amplimer obtained and sequenced from isolate AR13.1/ (Table 4) in which the region between Tn554 and the mec complex was absent but which carried a mec class A.3 mec complex. All isolates with

10 VOL. 49, 2005 NOVEL SCCmec VARIANTS AND mec CLASSES IN IRISH MRSA 2079 FIG. 3. Schematic diagram of the genetic organization of the mec complexes of staphylococci. The illustrations of mec complexes A, B, C1, C2, and D are based on studies by Katayama and coworkers (22), and the illustration of mec complex E is based on studies by Lim and coworkers (22, 24). The mec classes A, B, C2, and E indicated in bold have been previously identified in MRSA (18, 22, 24). The novel mec classes A.3 and A.4 were identified in the present study of Irish MRSA isolates. Class A.3 mec differs from class A mec in that it has a copy of meci that is truncated at the 3 end and an upstream copy of IS1182 (IS1182- meci-mecr1-meca-is431mec). meci is only 253 bp, whereas intact meci is 372 bp. In class A.4, mec IS1182 is present within meci and there is a 16-bp deletion of meci immediately upstream of IS1182 ( meci-is1182- meci-mecr1-meca-is431mec). the SCCmec II kdp (n 10) and SCCmec II kdp & meci (n 10) harbored the integrated plasmid pub110 and the second copy of IS431 found in SCCmec. Isolates with multiplex patterns SCCmec II kdp & pub110 (n 27) and SCCmec II kdp, meci & pub110/iv (n 7) gave the same size amplimer as that obtained from isolate AR13.1/ for amplification of the region across pub110 (Table 4), indicating the absence of pub110 and the second IS431 sequence.

11 2080 SHORE ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 6. Identities between ORFs of the novel SCCmec IV variant element IVE of Irish MRSA isolate AR43/ and the conventional SCCmec IVc element SCCmec element and ORF Location a Homology to ORFs of SCCmec IVc b ORF % Homology c Gene product IVE CR Hypothetical protein IVE CR Hypothetical protein IVE R Hypothetical protein IVE R Hypothetical protein IVE R Hypothetical protein IVE06 (ccra2) R Cassette chromosome recombinase A2 IVE07 (ccrb2) R Cassette chromosome recombinase B2 IVE R Hypothetical protein IVE R Hypothetical protein IVE R Hypothetical protein IVE R Hypothetical protein IVE R Hypothetical protein IVE CR Putative transposase of IS1272 IVE CR010 ( mecr1) 100 Truncated signal transducer protein MecR1 IVE R012 (meca) 100 Penicillin binding protein 2a IVE CR Hypothetical protein IVE R Transposase of IS431mec IVE d Hypothetical protein IVE Hypothetical protein IVE Hypothetical protein a The nucleotide positions were determined based on the sequences for SCCmec IVE deposited in the GenBank database under accession number AJ b ORFs and gene products from SCCmec IVc nucleotide sequence in the GenBank database, accession number AB (19). c Based on amino acid identity. d, no homology. Additional novel variants of SCCmec IV. All 20 isolates with the AR43 phenotype which were found to carry SCCmec IV by the simplex method but gave the multiplex pattern SCCmec IV dcs produced the expected size products for a type IVc SCCmec element when the primers used with isolate AR43/ were used, but differences were observed following amplification of the L-C and I-R regions (Table 4). Seventeen of the isolates had the same L-C region as SCCmec IVc (6.1-kb amplimer), but three isolates yielded a 5.6-kb product (Table 4). Sequencing of this amplimer confirmed it contained the SCCmec IVb L-C region. As with isolate AR43/3330.1, an amplimer that was larger than expected was obtained when the I-R region was amplified in all AR43 isolates (Table 4). This 4.7-kb region from an AR43 isolate with an L-C region similar to SCCmec IVb was sequenced and was 100% identical to that of isolate AR43/ DISCUSSION MLST and SCCmec element analysis of MRSA isolates recovered in Irish hospitals between 1971 and 2002 showed that clones representative of each of the five major clonal complexes and eight of the major pandemic lineages have been present in Ireland at some time over the past 30 years (Tables 1 and 3). This study also confirmed that there have been major changes in the dominant clonal types. The genotypes of the earliest MRSA during the 1970s and early 1980s were ST250- MRSA-I or its novel variant, ST250-MRSA-I pls, followed in the mid- to late 1980s by a new variant, ST239-MRSA-III pi258 & Tn554. In 1989, ST239-MRSA-III predominated, but during the 1990s this was displaced by ST8-MRSA-II, exhibiting six of the nine novel SCCmec variants identified in this study. In 2002, the dominant clone was ST22-MRSA-IV. Previous studies have proposed that MRSA clones should be defined on the basis of their genetic background as identified by MLST and their SCCmec type, but a striking finding of the present study was the extent of variation within SCCmec. If clones among the Irish MRSA population are defined on the basis of all SCCmec variants recognized, the population includes 20 clonal types consisting of 10 previously recognized clonal types, 9 with previously described genetic backgrounds but with novel SCCmec variants (including two new variants of class A mec) and one with a previously unreported genetic background (a double locus variant of ST5 in association with SCCmec II) (Table 3). The greatest variation was seen among isolates with the ST8 genotype. For example, among the 54 isolates exhibiting SCCmec IV by the simplex method, only those with the ST8 genotype (20/54) had variant SCCmec elements. In fact, 86% of isolates (74/86) with novel SCCmec variants had the ST8 genotype, and all carried previously unreported variants of SCCmec II (IIA to IIE) or IV (IVE and IVF). Because both SCCmec I and IV have a mec complex with IS1272 inserted at the same junction point, it has been suggested that recombination has occurred between SCCmec I and other sequences to generate SCCmec type IV (25). The novel ST8 SCCmec variants identified in this study resemble rearrangements of SCCmec II and SCCmec IV. SCCmec variants IIA to IIE have an L-C region that is almost identical to that of SCCmec IVb, but the rest of each element has closest identity with SCCmec II, including a very similar genomic structure (Fig. 2). SCCmec IIA to IIE appear to be closely related. SCCmec IIB was found in isolates from 1989, making it the earliest SCCmec II variant identified in this study. It harbors the original class A mec complex, but its similarity to SCCmec IVb in the L-C region make its provenance unclear. It is likely that SCCmec IIB was generated following loss of most of Tn554,